@fevzikorkutata I don't understand what you're saying.
That fix is in 1.17.0 so you shouldn't have run into that issue if you went straight from 1.0.3 to 1.17.0 as I recommended.
Okay, I disabled it, but even after updating SSO, I get this error in the browser after google OR normal email sign-in:
{"error":"Cannot read property '_csrf' of undefined"}
Also, when I updated SSO, I got these warnings again.
npm WARN package.json [email protected] No repository field.
npm WARN package.json [email protected] No repository field.
npm WARN package.json [email protected] No repository field.
By the way, how do I get back to the purple/lavender color scheme?
Coolio, that seems easy enough. Any idea what I can do about the email/SSO sign on error?
{"error":"Cannot read property '_csrf' of undefined"}
Okay, this reproduces it:
Try to log in to nutritionask.com using email, not SSO
It (usually/always) results in this error:
{"error":"Cannot read property '_csrf' of undefined"}
Does this mean anything? (from output.log)
^[[32minfo^[[39m: [Auth] Session qiMn3JE_1AWce13a2lK4IWtL logout (uid: 1)
node_redis: no callback to send error: ERR unknown command 'pexpire'
Also, the heart buttons on each post turned into blocks...I'm trying to decide whether to hack away at this or find another route -- like is there a way to just make the recent topics list the homepage, without using the custom homepage plugin?
Mayhaps @psychobunny could add "redis needs to be updated like so" that to the NodeBB for Dummies install guide? I just followed those instructions like a robot, and robots don't know what needs updating
Like now I'm trying to update redis and got this message when running the install script:
Starting Redis server...
*** FATAL CONFIG FILE ERROR ***
Reading the configuration file, at line 55
>>> 'tcp-backlog 511'
Bad directive or wrong number of arguments
ERROR: Failed starting service.... Aborting!
...and I still get this message when trying to log in to my forum using email+password, with the latest stable redis downloaded but not working apparently?
{"error":"Cannot read property '_csrf' of undefined"}
I just used the digitalocean guide to getting the newest redis, but changed the path to get the latest stable version. Is there any way to roll back to 0.4.2? My forum is all jacked up now!
haha woops. yeah that dummies guide really needs to be updated thanks @a_5mith
Nope, didn't set up a redis password.
The only thing I did is type in those four lines, check the version (successfully updated!) and then blammo the forum's gone now. I suppose I could revert to a backup or start all over, but I'd rather try to avoid that
These are the files:
drwxr-xr-x 2 root root 4.0K May 26 01:24 6379
-rw-rw---- 1 redis redis 1.1M May 26 14:26 dump.rdb
drwxrwxr-x 6 root root 4.0K Apr 22 04:14 redis-2.8.9
-rw-r--r-- 1 root root 1.1M Apr 22 04:15 redis-2.8.9.tar.gz
-rw-rw---- 1 redis redis 131K May 26 15:44 redis.rdb
The forum doesn't have anything on it after the redis update, and my admin login doesn't work. I haven't done anything except follow the digitalocean install guide, set up reverse proxy in nginx.conf, and update redis.
I think I know what happened...
old redis (< 2.4..?) had default database file as dump.rdb
. Newer redis configs have the database file set to redis.rdb
.
You should have a config file for redis at /etc/redis/redis.conf
. Look for a line that starts with dbfilename
. What is its value? It's probably redis.rdb
. Change it to dump.rdb
and save. Then restart redis: service redis-server restart
It was already dump.rdb
so I left it that way. Restarted redis and still no forum.
Could the directory structure be at fault? I didn't have the new redis at the location you pointed to, it was at /var/lib/redis/redis-2.8.9. (both redis.conf files point at dump.rdb by the way).
EDIT: It's back now! You da' man. I'm still not sure which is the real redis.conf file though.